Open Collections

Abstract

We are concerned with iterative solvers for large and sparse skew-symmetric linear systems. First we discuss algorithms for computing incomplete factorizations as a source of preconditioners. This leads to a new Crout variant of Gaussian elimination for skew-symmetric matrices. Details on how to implement the algorithms efficiently are provided. A few numerical results are presented for these preconditioners. We also examine a specialized preconditioned minimum residual solver. An explicit derivation is given, detailing the effects of skew-symmetry on the algorithm.